Your company's answers are buried across the systems you use every day. MerchantsBI brings that information together, gives AI the context to understand it, and puts the answers at your fingertips.
01ConnectConnect every system once, not once per AI tool.
01ConnectConnect every system once, not once per AI tool.
Each system needs its own MCP server: sign-in, rate limits, field mapping and hosting. When one system changes, that piece breaks and someone on your team fixes it.
Pick your systems from 56 connectors. MerchantsBI handles sign-in, limits and fields, and brings everything into one model the AI reads.
02SelectOnly the columns that hold data come across.
Every column is pulled on every run, including the ones that are always empty. The AI has to wade through all of them.
Columns are checked for fill before they are kept. In this NetSuite example, 151 of 342 made the cut.
03CleanCodes become names your team reads.
ORD_TYP R, RGN 04, CHNL 07. Someone decodes them in each report, and each report decodes them a little differently.
Columns get names your team reads, and functions run on each one while the table loads: trim and set the case of text, cast types, turn text into dates and numbers, and fill blanks. The original rows stay underneath.
04ModelJoins set up once and checked, so totals stay true.
Join on order number alone and two companies' orders collide. Add a style master with two rows per style and totals quadruple, with no error to warn you.
How your systems relate is set up once, with you, and checked before it saves, so totals stay true.
05DefineOne definition, used everywhere.
Finance takes out returns, sales doesn't, ops leaves out samples. Each number is defensible, and none of them match.
Net sales is defined once and reused across dashboards, reports and questions.
06TeachYour team's words, taught once.
Each new word, like doors, sell-through or the West, gets pasted into a prompt. It keeps growing and still misses the words nobody has added.
Teach a term once: what it means and what people call it. Words it can't place land in a review queue.
07RunPrepared ahead of time, not rebuilt on every question.
Every question pulls, joins and cleans from scratch, for every person who asks. A complex question can take many tool calls across several systems before an answer comes back.
Data is prepared in a batch run, so questions read data that is already joined and cleaned.
08ServeAnswers come from MerchantsBI, not your live ERP.
Every dashboard click and question becomes a query on your live ERP, competing with the people running the business on it.
Dashboards, reports and questions run on MerchantsBI's copy of your data. Your ERP is read in batches.
If you sell in stores, online and wholesale, your numbers live in separate systems. Someone on your team turns them into answers: the ops manager, the analyst, finance.
Every new question, like why margin dropped in the West, starts over. With spreadsheets, someone exports, combines and checks again while the decision waits. With AI connected to each system, every question means minutes of paging through raw records, matching products across systems and working out the formula again, for every person who asks.
Either way, the work behind one answer doesn't carry over to the next.
MerchantsBI helps your team answer the next question without starting over.
“Why did margin drop in the West?”
Gross margin in the West fell 3.3 points in October, to 37.9%. Most of the drop came from wholesale, where two accounts got deeper discounts.
| Channel | Sept | Oct |
|---|---|---|
| Wholesale | 38.4% | 32.3% |
| Retail stores | 44.1% | 42.1% |
| Online | 46.0% | 45.2% |
Gross margin in the West fell 3.3 points in October, to 37.9%. Most of the drop came from wholesale, where two accounts got deeper discounts.
| Channel | Sept | Oct |
|---|---|---|
| Wholesale | 38.4% | 32.3% |
| Retail stores | 44.1% | 42.1% |
| Online | 46.0% | 45.2% |
Gross margin in the West fell 3.3 points in October, to 37.9%. Most of the drop came from wholesale, where two accounts got deeper discounts.
“Is that wholesale or our own stores?”
Wholesale explains about 71% of the drop; your own stores about 22%. Online is a small part of it.
| Channel | Share of the drop |
|---|---|
| Wholesale | 71% |
| Retail stores | 22% |
| Online | 7% |
Wholesale explains about 71% of the drop; your own stores about 22%. Online is a small part of it.
| Channel | Share of the drop |
|---|---|
| Wholesale | 71% |
| Retail stores | 22% |
| Online | 7% |
Wholesale explains about 71% of the drop; your own stores about 22%. Online is a small part of it.
“What was sell-through last month, across all channels?”
Sell-through for October was 48.1% across all channels, down from 51.6% in September. Retail stores are pulling the total down: the most inventory and the lowest rate.
| Channel | Units sold | Units available | Sell-through |
|---|---|---|---|
| Online store | 4,820 | 9,600 | 50.2% |
| Retail stores | 6,150 | 14,300 | 43.0% |
| Wholesale | 2,400 | 3,900 | 61.5% |
| Total | 13,370 | 27,800 | 48.1% |
I calculated sell-through as units sold ÷ (beginning inventory + units received). Let me know if you count returns differently.
Sell-through for October was 48.1% across all channels, down from 51.6% in September. Retail stores are pulling the total down: the most inventory and the lowest rate.
| Channel | Units sold | Units available | Sell-through |
|---|---|---|---|
| Online store | 4,820 | 9,600 | 50.2% |
| Retail stores | 6,150 | 14,300 | 43.0% |
| Wholesale | 2,400 | 3,900 | 61.5% |
| Total | 13,370 | 27,800 | 48.1% |
I calculated sell-through as units sold ÷ (beginning inventory + units received). Let me know if you count returns differently.
Sell-through for October was 48.1% across all channels, down from 51.6% in September. Retail stores are pulling the total down: they carry the most inventory and had the lowest rate.
| 👁 | Channel | Units Sold | Units Available | Sell-Through |
|---|---|---|---|---|
| 1 | Online store | 4,820 | 9,600 | 50.2% |
| 2 | Retail stores | 6,150 | 14,300 | 43.0% |
| 3 | Wholesale | 2,400 | 3,900 | 61.5% |
| 4 | Total | 13,370 | 27,800 | 48.1% |
“Which categories are dragging it down?”
Running and Trail are dragging retail sell-through down; Apparel is above average. Running carries the most retail inventory.
| Category | Sell-through |
|---|---|
| Running | 39.0% |
| Trail | 41.5% |
| Lifestyle | 46.2% |
| Apparel | 52.8% |
Using the same sell-through formula as before.
Running and Trail are dragging retail sell-through down; Apparel is above average. Running carries the most retail inventory.
| Category | Sell-through |
|---|---|
| Running | 39.0% |
| Trail | 41.5% |
| Lifestyle | 46.2% |
| Apparel | 52.8% |
Using the same sell-through formula as before.
Running and Trail are dragging retail sell-through down; Apparel is above average. Running carries the most retail inventory.
“Did the Amazon promo pull sales from our site?”
Site sales dipped 6% during the Amazon promo week and recovered the week after. No lasting shift from your site to Amazon.
| Week of | Site sales | Amazon sales |
|---|---|---|
| Oct 6 | $182k | $96k |
| Oct 13 (promo) | $171k | $148k |
| Oct 20 | $185k | $101k |
Site sales dipped 6% during the Amazon promo week and recovered the week after. No lasting shift from your site to Amazon.
| Week of | Site sales | Amazon sales |
|---|---|---|
| Oct 6 | $182k | $96k |
| Oct 13 (promo) | $171k | $148k |
| Oct 20 | $185k | $101k |
Site sales dipped 6% during the Amazon promo week and recovered the week after. No lasting shift from your site to Amazon.
Connect. Connect your systems in a few clicks and see every table they hold.
Build action plans. Turn raw tables into clean, joined, report-ready data with step-by-step action plans.
Schedule. Decide when each system refreshes, from hourly to nightly, inside the hours you allow.
Sync. Watch every sync land, system by system and table by table, with row counts and times.
Navigation. Arrange the pages each team sees and how each one opens.
Model relationships. Define how your tables join, once, and check every join before it goes live.
Data tagging. Teach MerchantsAI the words your team uses for every column and code.
Formula builder. Write a measure once, like net sales or sell-through, and use it everywhere.
Dashboards. Give every team a live board where one filter updates every tile.
Reports. Build pivot reports by dragging fields, then format, drill into and export them.
MerchantsAI. Ask a question in plain language and get the answer, the chart and the table behind it.
Scheduled reports. Send saved reports to the right people on a schedule, as Excel or CSV.
Alerts. Coming soon: get told when a number moves, with the rows behind it.
Source Selection
Column Mapping
Source profiled 9 Nov 14:02 · 412,880 rows · 151 columns have values · 191 are empty Re-profile
Review Summary
A report page is built on one activated table.
IconSelect a relationship on the left to inspect it.
| sales_order_lines | = | items | |
|---|---|---|---|
| item_id | = | id | ✕ |
| No join columns — nothing can join without them. | |||
Columns pair row by row.
+ Column pairmany-to-one — safe for totals
07198,18248%03127,99331%online = 031170,19017%stores = 110516,5154%| Jul | Aug | Sep | |
|---|---|---|---|
| Running | 44% | 43% | 41% |
| Trail | 38% | 33% | 31% |
| Lifestyle | 39% | 40% | 38% |
| Apparel | 16% | 14% | 15% |
| Order | Date | Style | Channel | Store | Units |
|---|---|---|---|---|---|
| SO-48211 | Sep 28 | Pegasus 41 | Retail | Denver | 24 |
| SO-48207 | Sep 28 | Vomero 18 | Online | — | 2 |
| SO-48190 | Sep 27 | Structure 26 | Wholesale | — | 180 |
| SO-48176 | Sep 27 | Trail Ridge 3 | Retail | Boulder | 11 |
| SO-48162 | Sep 26 | Pegasus 41 | Amazon | — | 36 |
| SO-48151 | Sep 26 | Summit Knit | Retail | Austin | 8 |
| 2025 · Q2 | 2025 · Q3 | |
|---|---|---|
| Category | Sales | Sales |
| ▸ Footwear | 412,180 | 455,020 |
| ▸ Apparel | 206,340 | 231,510 |
| Total | 618,520 | 686,530 |
| 2025 · Q2 ▾ | 2025 · Q3 ▾ | |||||
|---|---|---|---|---|---|---|
| Category ▸ Sub ▸ Style | Sales | Units | Margin ↓ | Sales | Units | Margin |
| ▾ Footwear | 412,180 | 18,412 | 38% | 455,020 | 19,904 | 41% |
| ▾ Running | 228,340 | 9,116 | 42% | 241,760 | 9,842 | 44% |
| Pegasus 41 | 96,410 | 3,388 | 47% | 104,220 | 3,702 | 48% |
| Structure 26 | 60,908 | 3,114 | 44% | 68,600 | 3,722 | 45% |
| Vomero 18 | 71,022 | 2,614 | 31% | 68,940 | 2,418 | 29% |
| ▾ Trail | 183,840 | 9,296 | 33% | 213,260 | 10,062 | 37% |
| Trail Ridge 3 | 102,400 | 5,180 | 35% | 118,120 | 5,560 | 39% |
| Summit Trail 2 | 81,440 | 4,116 | 31% | 95,140 | 4,502 | 35% |
| ▾ Apparel | 206,340 | 11,208 | 14% | 231,510 | 12,640 | 16% |
| Summit Knit | 92,110 | 4,320 | 18% | 104,860 | 4,910 | 19% |
| Tempo Short | 71,480 | 5,102 | 12% | 80,230 | 5,740 | 15% |
| Ridge Jacket | 42,750 | 1,786 | 9% | 46,420 | 1,990 | 12% |
| Σ Grand total | 618,520 | 29,620 | 30% | 686,530 | 32,544 | 33% |
| Order | Date | Customer | Channel | Units | Net sales |
|---|---|---|---|---|---|
| SO-47702 | Sep 24 | Harbor Stride | Wholesale | 240 | $6,912 |
| SO-47688 | Sep 23 | Online store | Online | 3 | $420 |
| SO-47651 | Sep 22 | Denver store | Retail | 18 | $2,394 |
| SO-47630 | Sep 21 | Amazon | Amazon | 41 | $5,084 |
| SO-47609 | Sep 20 | Ridgeline Run Co. | Wholesale | 120 | $3,456 |
Gross margin in the West fell 3.3 points in October, to 37.9%. Most of the drop came from wholesale, where two accounts got deeper discounts.
Wholesale explains about 71% of the drop; your own stores about 22%. Online is a small part of it.
Sell-through for October was 48.1% across all channels, down from 51.6% in September. Retail stores are pulling the total down: they carry the most inventory and had the lowest rate.
| 👁 | Channel | Units Sold | Units Available | Sell-Through |
|---|---|---|---|---|
| 1 | Online store | 4,820 | 9,600 | 50.2% |
| 2 | Retail stores | 6,150 | 14,300 | 43.0% |
| 3 | Wholesale | 2,400 | 3,900 | 61.5% |
| 4 | Total | 13,370 | 27,800 | 48.1% |
Running and Trail are dragging retail sell-through down; Apparel is above average. Running carries the most retail inventory.
MCP is how an AI assistant reaches your data. MerchantsBI does the heavy lifting behind it: it brings your systems together, joins them, applies your formulas and permissions, and serves it all to your AI through one MCP connection.
Use MerchantsAI inside MerchantsBI, your own AI over MCP, or both. They work from the same model.
Bring your systems together, ask the next question, and get an answer your whole team can work from.